masterThesis
Análise estatística de métodos de interpolação espacial e técnicas de sensoriamento remoto usando imagem de satélite na caracterização batimétrica da Bacia Potiguar
Fecha
2018-04-20Registro en:
CAMPOS, Pedro Henrique Pinto. Análise estatística de métodos de interpolação espacial e técnicas de sensoriamento remoto usando imagem de satélite na caracterização batimétrica da Bacia Potiguar. 2018. 87f. Dissertação (Mestrado em Engenharia Civil) - Centro de Tecnologia, Universidade Federal do Rio Grande do Norte, Natal, 2018.
Autor
Campos, Pedro Henrique Pinto
Resumen
The extraction of bathymetry by satellite image is an activity that has been evolving in
its methodology due to the aerospacial technological advances. Satellite images of a
wide electromagnetic capture are able to bring information that are beyond traditional
photographs. They are images that, through some specific calculations methodologies
can show, for example, the depth of the ocean. The literature shows that it is not yet
known which calculation methodology is most appropriate. Another key point involved
in the process is the generation of bathymetric maps by interpolation from clouds of
points, there are several methods of interpolation today, but much is still debated about
which are the most appropriate for each case. The mapping of the seabed is an activity
performed for several purposes. It can be used for navigation, oil reservoir prospecting,
environmental monitoring, wave modeling, coastal erosion study and others. Brazil has
one of the largest coastlines in the world with more than 8500 km of extension, many
of its cities have been installed on the coast and are submitted to the consequences
of problems associated with the sea, coastal erosion, urban disorder, landscape
impact, ecological impact etc. This necessitates a deeper study of ocean floor mapping
models. Moura (2010) did the interpolation analysis using bathymetry data obtained by Nautical Charts and bathymetry data obtained in loco. In his work a series of 12
interpolators were evaluated. A satellite image of Gomes et al. (2007) was used to aid
in a visual evaluation of the results. The present work has as general objective to
complement the analysis of Moura (2010), providing a quantitative reference to make
this evaluation, from a bathymetry extraction of a Landsat 8 Satellite Image of the
region. The methods of bathymetry extraction used were Mcfeeters (1996) and Philpot
(1989). The extraction of bathymetry was done with the software Arcgis, Envi and
Matlab. The interpolations were made in the Surfer and the data were compared to
each other using the statistical analysis of Correlation Coefficient (R²), concordance
index (d), MAPE, RMSE and Cross Validation. In terms of the results of the bathymetry
extraction, Mcfeeters (1996) was the method that brought better representation of the
region. In terms of result of the interpolations of the in-situ bathymetry data, the
Interpolators Local Polynomial, Inverse Distance, Minimum Curvature were chosen the
most representative methods. For nautical chart data, the Natural Neighbor,
Triangulation with Linear Interpolation, Kriging, and Radial Base Function interpolators
were classified as the most appropriated interpolation methods because they had the
best statistic results.